Rukmini Sai Rupa Sri Mada

This is one of the warnings i am getting during the simulation

 The eigenproblem is ill-conditioned (relative residual from test linear solve: 6.084859E-03).  The accuracy of the solution may not be satisfactory.  Modifying the initial value of the frequency shift  (FREQB in the MODOPT command) or changing the boundary conditions of the model might solve the problem.




How do I address this?